On July. 12, Elon Musk announced the launch of a new artificial intelligence company: xAI. The company’s website claims that it’s led by Elon Musk, CEO of Tesla and SpaceX, and has a goal “to understand the true nature of the universe.”
xAI further clarifies that it is “a separate company from X Corp, but will work closely with X (Twitter), Tesla, and other companies to make progress towards our mission.” It was built by talents who previously worked for DeepMind, OpenAI, Google Research, Microsoft Research, and Tesla.
In a Tweet posted on the same day of xAI’s launch, Musk revealed his reason for choosing July 12, 2023, to launch the program: it is a direct nod to The Hitchhiker’s Guide to the Galaxy by Douglas Adams.
He explained that 7+12+23=42. In Adams’ book, a supercomputer named Deep Thought is asked to find the “Answer to the Ultimate Question of Life, The Universe, and Everything,” to which it famously replies: “42”.
xAI’s co-founder Greg Yang tweeted on July 12 that “developing the ‘theory of everything’ for large neural networks will be central to taking AI to the next level.” He continued, explaining it “will enable everyone to understand our mathematical universe in ways unimaginable before.”
Both Musk and Yang retweeted an invitation to an xAI Twitter Space Chat (Friday, July 14, PDT) for a meeting with the xAI team members.
xAI’s first info session
The info session started 13 minutes late than scheduled. Musk started the session by explaining they needed to “tweak the algorithm” to promote the chat to more users.
During the 90-minute Space session, over 30,000 listeners attended. Musk revealed that he quietly started xAI in April for a formal entry into the AI market. The ultimate goal of the enterprise is to “understand the true nature of the universe,” Musk said.
The overarching goal of xAI is to build a good AGI [artificial general intelligence] with the overarching purpose of just trying to understand the universe.Elon Musk
Artificial general intelligence (AGI) has long sparked debate among AI experts. The safety and morality of AI is at the center of the debate. Elon Musk has been very vocal about this potential threat.
The billionaire signed an open letter in March calling for more regulation over the technology. He signed alongside 30,000 other people, including Apple co-founder Steve Wozniak and many other tech dignitaries.
In the Twitter Space session earlier today, Musk claimed that his new artificial intelligence company xAI could be better at building safe AI systems than the existing players.
He blamed Open AI for becoming “voracious for profits” and hinted that “all these ESG [environmental, social and governance] mandates and stuff that push (public) companies (like Google or Microsoft) in questionable directions.”
He claimed that “[A]s a company that’s not publicly traded, xAI is not subject to market-based incentives, or the non-market-based, ESG incentives.” xAI, on the other hand, is “a little freer to operate.”
Twitter data training expected
The Tesla CEO admits that xAI will use Twitter data to train its artificial intelligence systems and products. Musk did not specify whether or how much Twitter will charge xAI or other companies for its data.
Only one week ago, Musk placed rate limits on Twitter users “to address extreme levels of data scraping & system manipulation.” He made this decision despite extreme controversies. He said, “We had multiple entities scraping every tweet ever made, and trying to do so in, like, basically a span of days.”
During the Friday xAI info session, Musk alleged that “Every AI organization on Earth” had used Twitter’s data for training. Musk failed to provide any evidence for this statement. Earlier this same month, Twitter sued four unknown parties in Texas for data scraping.
Collaboration with other Elon Musk companies
The Tesla CEO admitted that the xAI startup will collaborate with the Tesla auto company, not only on the “silicon front,” but also on the “AI software front.” Musk denied that Tesla is building anything like a GPU (Graphic Processing Unit) or an equivalent of a GPU.
Musk also mentioned Dojo, a supercomputer Tesla is developing for AI machine learning and computer vision training. Tesla is known for using video clips and data from its customers’ vehicles to improve existing software and to develop new features.
The CEO concluded that the eventual xAI language model will not be a “politically correct” one. “I think our AI can give answers that people may find controversial even though they are actually true,” he said.